Can the SKEW/VIX predict market corrections?

Over the past few years, we have developed a series of indicators of market complacency in order to prevent investors of a sudden spike in price volatility after a long period of calm. One of them looks at the divergence between the Economic Policy Uncertainty EPU index (Baker et al., 2016), a measure of economic uncertainty based on newspaper coverage frequency, and the VIX. Figure 1 (right frame) shows that over the past few years, the EPU index has been displaying a much higher risk level that would be inferred from the options market. Some also watch the activity in the TED spread, which has been distorted since the financial crisis due to a tightening up of regulations and changes in money market funds (figure 1, right frame), and notice when it starts to stir.

Figure 1

Fig1 New

Source: Eikon Reuters

An interesting one looks at the behavior of the SKEW index relative to the VIX. As we previously mentioned, since the crash of October 1987 (Black Monday, DJ down 22.6%), investors have realized that S&P500 tail risk (returns that are under 2 or more standard deviations below the mean) is significantly greater than under a log-normal distribution. Therefore, the ‘skew’ measures the perceived tail risk of the market via the pricing of OTM options. A rise in skew indicates that ‘crash protection’ is in demand among institutional investors. A ‘low VIX/high skew’ combination says that the market is complacent, however the ‘big players’ perceive far more tail risk than usually. Therefore, a surprise increase in realized volatility may not be too far away.

Figure 2 shows the times series of the VIX, the SKEW and the weekly change in the SP500. As you can see, a sustain period of falling VIX and rising SKEW is generally followed by a sharp spike in implied volatility. For instance, while the VIX was approaching the 10 level in the summer of 2014, the SKEW index had been on a rise in the year preceding that period, rising from 113 in July 2013 to 146 in September 2014. In addition, the Fed stepped out of the bond market in October 2014 (end of QE3) and therefore exacerbated investors’ concern on the market. We then saw huge moves in both equities and bonds in the middle of that month and the 18-month period that followed was basically a flat equity market with some significant drawdowns (October 2014, August 2015 and December 2015 / January 2016). The second period we highlighted was in 2016 / 2017, which was marked by an extremely low volatility and a rising SKEW. We saw that things reverted drastically in the February VIX-termination event. Following this event, we eventually had another period of falling VIX and rising SKEW in the next months before the October sell-off. We can notice in the chart that the SKEW does not stay above the 150 threshold for too long, hence a VIX trading at around 12 and a SKEW at 150 were last summer were indicating a potential market turmoil.

Figure 2

Fig2 New (1)

Source: Eikon Reuters

The SKEW/VIX behavior does not predict a market correction all the time (i.e. the SKEW had been falling for months prior the August 2015 sell-off), however we think that investors should remain cautious when the SKEW starts to rise above 140 and the VIX remains low. While a falling VIX would push investors to increase their leverage (target vol strategies or risk parity funds), we think that looking at the two variables for portfolio construction could help reduce the potential drawdowns.

FX Cross-Currency Basis Swaps and Hedging Costs

One interesting topic in the FX market that has been closely studied by both academics and practitioners over the past decade is the violation of the covered interest parity (CIP). CIP is a textbook no-arbitrage condition that states that interest rate differential between two currencies should equal the differential between the forward and spot exchange rates. In other words, in discrete time, we have the following condition:

Where S is the spot exchange rate, F is the forward exchange rate, i is the domestic interest rate and i* is the foreign currency interest rate. The problem is that the above equation has held since the Great Financial Crisis; as it started to become more expensive to borrow US Dollars against most currencies during periods of stress, the cross-currency basis swap (CCBS) has been diverging from zero for the Euro, the British pound and the Japanese Yen. Figure 1 (left frame) shows the evolution of the 3-month CCBS for the three currencies (against the USD) since 2012.

Low interest rates combined with the significant divergence in unconventional monetary policies run by the BoJ and ECB over the years has put pressure on the exchange rates and the CCBS, and therefore has increased the hedging costs for Euro and Japanese investors. The current rate on the US 10Y Treasuries (3.15%) looks certainly very interesting for unhedged international investors (relative to domestic bonds such as in the Euro area or Japan), however changes drastically when we adjust for hedging costs. Figure 2 represents the cash-flows that occur at the start, during the term and at maturity when a Euro investor (A) enters a cross currency basis swap. As you can see, each quarter A pays the 3M USD Libor and receives the 3M Euribor and the basis. Hence, the more negative is the basis, the higher the hedging cost. With the 3M Euribor at -0.316%, the 3M CCBS at -44bps and the 3M Libor USD at 2.61%, the current return on a FX-hedged 10Y US Treasuries is negative (-20bps). Figure 1 (right frame) shows that despite the rise in US yields since the middle 2016, it has been falling for Euro and JPY investors after adjusting for FX hedging costs. A UK investor would get an annual return of 1.35%, which is 15bps below he can get in holding a 10Y Gilt.

Figure 1


Source: Eikon Reuters


Figure 2


Source: BIS


Great Chart: AUDJPY vs. EM equities

Prior the Financial Crisis, the carry trade strategy in the currency market was perceived to be a profitable and generated significant returns for traders seeking for yields. In the appendix A, we show the performance of the carry strategy between 1975 and 2008, along with the performance of equities and fixed income according to a 2008 publication from JP Morgan. We can notice that the funded carry strategy, which invested equally in three currencies with the highest yields funded by borrowing from the three currencies with the lowest yields, outperformed both fixed income and equity returns during that period. According to JPM calculations, if you invested $1 in 1975 in each of the strategy, the initial investment in funded carry grew to $84.16 in early 2008 (vs. $15.25 in fixed income and $51.74 in equities), whilst experiencing volatility levels between those two assets.

However, the situation changed abruptly during the financial crisis when the carry currencies (i.e. AUD) plummeted and funding curries (i.e. JPY) experienced significant appreciation. For instance, if we take the AUDJPY exchange rate as a proxy of the traditional G10 carry trade, we can see in Appendix B that the strategy followed the same pattern as the (US) equity market and hence experienced a sharp correction between July 2008 and March 2009 (AUDJPY was down 45%). Since then, many investors have considered the carry trade strategy to be a risk-on strategy, exhibiting strong co-movements with DM equity markets (hence poor for diversification) and described it as a ‘gradual appreciation punctuated by sudden crashes’ type of behavior (the famous quote: ‘going up by the stairs, and coming down by the elevator’). It is quite usual for an EU/US global macro trader or investor to watch the overnight Yen developments to see if anything major happened in Japan or China for example (strong Yen appreciation usually means bearish macro news for equities).

Even though the co-movement between AUDJPY and US equities (SP500) has been inexistent over the past 5 years, an interesting observation emerges when we overlay the AUDJPY exchange rate with EM equities. As you can see it on the chart, EM equities have moved in tandem with the ‘carry’ exchange rate; the 3M daily realized correlation stands now at 92%. AUDJPY is almost down 10 figures (i.e. 12%) since mid-January, and traded below 79 earlier this month, its lowest level in two years. We will see if the correlation persists in the months to come and if a rebound in the Aussie (or Yen weakness) will benefit to EM equities, which are down more than 25% since January highs.

Chart: AUDJPY vs. EM Equities (Source: Eikon Reuters)


Appendix A: Carry Strategy vs. Equities and FI (Source: JP Morgan)Carry.PNG

Appendix B: AUDJPY vs. SP500 (Eikon Reuters)AUDJPY 2008.PNG

How long will the UK equity market hold?

With an expected annual real growth of 1.5% in 2018 according to the general consensus, the UK economy switched from the top DM performer in 2014 to bottom 2 in 2018, the second lowest growing country after Japan (1.2%).  Uncertainty around Brexit is still weighing on the economic outlook, which is pushing the Bank of England to keep a loose monetary policy and use different tools to counter a potential downturn. As a result of the referendum, policymakers announced in August 2016 a GBP 70bn expansion of the central bank’s balance sheet, purchasing GBP 60bn of government bonds (Gilts) and GBP 10bn of corporate bonds. In addition, the BoE also provided GBP 127bn of cheap loans (TFS drawings) to banks through the Term Funding Scheme (TFS), a 4-year funding at the BoE Base Rate plus a fee to the banks requiring them to lend into the real economy (an equivalent to the LTROs in the Euro area). Therefore, if we combine the three outstanding amounts together (figure 1, left frame), the BoE balance sheet’s total assets currently stand at GBP 572bn, or roughly 28% of UK’s GDP. In addition, UK policymakers kept the Official Bank rate at low levels and have slowly started a tightening cycle (figure 1, right frame)The base rate currently stands at 0.75%, and market participants are pricing in 1 to 2 hikes by the end of 2019, with the Short-Sterling Dec19 futures contract trading at 98.85 (1.15% implied rate).

Figure 1


Source: Bank of England, Eikon Reuters

Even though the headline inflation in the UK is running at 2.5%, significantly helped by Sterling weakness following the referendum and the recovery in energy prices, policymakers are in a difficult position as the economic activity seems to be slowing down according to the fundamentals. Our leading economic indicator, which is built using a combination of survey and price data as inputs, is pricing a slowdown in industrial production within the next 6 months. Therefore, using the industrial production as a proxy of the UK’s economy activity, the real GDP annual growth could be actually be lower than the 1.5% expected for 2018.

Inflation should remain high according to our 6-month forecasting model, averaging an annual growth (CPI) of 2.5%/3% for the rest of the year, therefore nominal GDP in the UK should average 4% (which is far significantly higher than the 1.5% 10Y yield). Overall, political uncertainty around Brexit and the Trade war in addition to slowing fundamentals should limit growth expectations to the upside in the medium term and therefore should be reflected in the real and financial economy.

Figure 2


Source: Eikon Reuters, RR

Another interesting observation is the dramatic decrease in excess liquidity, computed as the difference between the annual growth of real M1 (CPI adjusted) and annual growth in industrial production. According to many empirical studies, an increase in excess liquidity should benefit to the risky assets such as the stock market. As you can see it in figure 3 (left frame), excess liquidity has significantly decreased from 10.15% in August 2016 to 0.51% in June 2018 and therefore could weigh on UK equities in the medium term. In figure 3 (right frame), we use excess liquidity as a 6M leading indicator that we overlay with the annual performance of UK financials (using Eikon Reuters Total Return Index), a sector which is usually considered to act as a barometer of the country’s economy. We can clearly notice that financials tend to perform badly in periods of decelerating excess liquidity.

Figure 3


Source: Eikon Reuters, RR

With the 10-year on Gilts trading slightly below 1.5% and an equity market up 1,500 points since Brexit (trading at 7,630 and 230pts away from the all-time high reached on May 21st), the British pound was the main asset that suffered from the Brexit vote. Cable plummeted from 1.4750 in early May 2016 to hit a low of 1.20 in October 2016 (its lowest level since 1985) before starting its recovery to 1.44 on the back of a US Dollar weakness in 2017. This year, Sterling is once again under pressure since the start of the Dollar rally in April, down 16 figures and currently trading below 1.29. Market sentiment is extremely bearish, with speculative investors net short -72.3K contracts according to the CFTC (August 21st CoT report). In figure 4 (left frame), we can notice that the 33K increase in longs was offset by the massive increase in shorts from -82.4K to -140.4K (-58K) over the past month. We think there is still room for GBP weakness in the next three months to come ahead of Brexit negotiations, but the premium and the convexity on shorting the pound at these levels are not that interesting in our opinion.

However, it seems that the equity market has not been reacting neither to Brexit uncertainty and the recent slowdown in UK fundamentals. Figure 4 (right frame) shows that over the past two years, to the exception of the early 2018 (global) equity sell-off, the Footsie 100 index has been significantly sensitive to a move in Sterling (see more here). For instance, Cable’s weakness starting in mid-April has helped pushed UK equities to hit new all-time highs, with the index soaring from 6,890 On March 26th to 7,860 on May 21st. Even though we may see some further GBP weakness in the months to come that could push UK equities to new highs, we think that current low levels of implied volatility (FTSE 100 VIX is currently trading at 11) offer a good opportunity for investors to hedge against a sudden sell-off within the next 6 months.

Figure 4


Source: Eikon Reuters, CFTC

Great Chart: US ST Yield Curve vs. Cyclical-Defensive Stocks

Lately, the sharp revision of the US annual saving rate (up 1.6% on average since 2010) shifted growth expectations to the upside and lowered the bottom of the unemployment rate for the next few quarters. For instance, Goldman revised its GDP growth to 3% in Q4 (from 2.5% previously) and to 2% for 2019 (vs. 1.75%) and expects the unemployment rate to bottom at 3% in 2020. As a result, some investors are starting to consider that we may see more rate hikes by the Fed than currently expected. With two more hikes priced in for this year and another two to three for 2019, market participants expect the Fed Funds rate to hit [at most] 3.25% by the end of next year, which is more or less in line with the  Fed’s dot plot released at the June meeting (median projection at 3.125% for 2019).

However, we saw that the market is not expecting any more hikes post-2019, which could be interpreted as the end of the tightening cycle by US policymakers. According to the Eurodollar futures market, the December 2019 and December 2020 implied rates are trading equally at 3.06%, which suggests that the US economic outlook is expected to slow down at the end of next year. Hence, an interesting analysis is look at which sectors should perform well within the next 12 to 24 months if we stick with the scenario that economic uncertainty will increase at the end of 2019. A classic strategy looks at the Cyclical vs. the Defensive stocks. The main difference between Cyclical and Defensive stocks is their correlation to the economic cycle; Cyclical stocks tend to do well in periods of economic expansion (relative to Defensive stocks) but tend to experience more losses during recessions. According to empirical research, one of the main aspects that drive Cyclical and Defensive stocks’ performance is the beta of these stocks (also called the market risk premium). As the Defensive stocks are more resilient to an economic downturn, their beta is lower than 1 (resp. higher than 1 for Cyclical stocks).

Therefore, if we take the EuroDollar (ED) Dec19-Dec20 implied rate yield curve as our leading indicator of the business cycle, a flattening yield curve should benefit to the Defensive stocks (vs. Cyclical stocks). However, the chart below tells us a different story (Original Source: Nomura). We looked at the relationship between Cyclical-versus-Defensive sectors and the Dec19-Dec20 ED yield curve since the summer of 2008, and noticed that the two times series have been diverging for the past two years. The yield curve has constantly been flattening during that period, however Cyclical stocks have outperformed Defensive Stocks. We chose Materials, IT and Industrials sectors for the Cyclicals and HealthCare, Telecom and Utilities sectors for the Defensives (Source: Thomson Reuters Total Return Indices), and compute the ratio of the Cyclicals and Defensives new indices (find attached the file).

If you expect the two series to convergence back together, this would imply either a sudden steepening of the yield curve or Cyclicals to underperform Defensives.

Chart: ED Dec19-Dec20 yield curve vs. Cyclical-Defensive stocks (Source: Eikon Retuers)

Sectors vs. YC.PNG

EXCEL DATA LINK ====> Sectors

Great Chart: US 2Y Yield vs. CFTC Specs Positioning

Over the past few weeks, we have noticed an interesting development in the US interest rate market. Since the beginning of the year, the 2-year interest rate has constantly been increasing on the back of a tightening monetary cycle ran by US policymakers; it is now trading at 2.65%, up from 2% in early January. However, the CFTC Commitment of Traders report shows that speculators have not followed the trend and have been reverting their positioning since January (as opposed to the 10Y positioning). As we can see it on the chart, total net specs positions increased from -329K contracts on January 30th to -19K last week (July 17th) on the back of a sharp reduction in shorts from -735K to -492K and some increase in longs (+67K), hence completely diverging from the 2Y yield. What generated this sudden reversal? We think that the rise in the 2Y may be done and that the short end of the curve could settle around the current levels for a while.

After two hikes this year, the Fed Funds rate currently stands at 2%, and another two moves are expected according to market participants for the rest of 2018 (at the September and December meetings). With the US economy expected to grow at 4.5% in real terms in the second quarter according to GDPnow forecasts, US policymakers have benefited from strong momentum in US fundamentals and a dull summer market with the 10Y yield trading quietly below 3% and equities steadily recovering from their February lows. The SP500 index is up 300pts from its low reached on Feb 9th, and currently trades 50pts away from its all-time high reached in January prior the equity rout. Even though financial markets have still got to ‘face’ the August low-liquidity period, which is the most volatile month if we look at the past twenty years, US policymakers have got all the conditions not to disappoint market participants in the following FOMC meetings.

However, we think that much of the action concerning US monetary policy has been priced in by investors, and we can’t see any more hawkish surprises coming in the following months. Therefore, the 2Y may stabilize around its current level at 2.7%, which could explain the reversal in the specs positions. It will be interesting to see how the short-end and the long-end of the curve react to a sudden rise in uncertainty by the end of the summer, pushing down drastically the probability of a hike at the September meeting.

Chart. 2Y US yield vs. CFTC Specs Positioning (Source: Reuters Eikon, CFTC)

Net specs.PNG

Great Chart: Oil Prices vs. Japan Trade Balance

The recovery in oil prices since February 2016 has eased financial conditions for most of the Middle East countries and has reversed the path of the corporate default rate for US energy companies exposed to the shale industry. Higher oil prices have also brought back inflation in most of the economies, hence pushed up expectations of nominal growth rates. However, for countries that are heavy importers of energy (i.e. Japan), higher oil prices usually mean a deterioration of the Trade Balance. Japan has limited domestic proved oil reserves (44 million barrels), which means that the country is a net importer of oil. According to the EIA, Japan is the fourth-largest petroleum consumer and the third largest net importer, and its daily consumption in 2016 was of 4 million barrels per day. Therefore, if we plot the WTI futures prices (6M lead) with the Japanese trade balance, we can notice a significant co-movement between the two times series. This chart suggests that oil prices can be used as a sort of leading indicator for the Japanese trade balance. For instance, when oil prices entered a bear market in 2014, the trade balance switched from a 1.1tr JPY deficit in the middle of 2014 to a 350bn JPY surplus in H2 2016. Hence, with oil prices constantly trending higher with the front-month contract on the WTI trading at $70 per barrel, its highest level since Q4 2014, we can potentially anticipate that the Japanese trade balance will go back into deficit in the medium term.

What are the consequence for the Japanese Yen?

In our BEER FX model, we saw that exchange rates (in log terms) react positively to a positive change in interest rate differential and in terms of trade differential, and negatively to a change in inflation rate differential. Hence, if we expect import prices to rise in Japan due to higher energy costs (especially Oil), the terms of trade should ‘deteriorate’ and therefore have a negative impact on the currency. However, we know that the Japanese Yen is also very sensitive to the current macro environment and often acts as a safe-have asset when the risk-off sentiment rises (Yen appreciates in periods of equity sell-off). In our view, the problem Japanese officials may face in the following 6 months is higher energy prices combined with a strong Yen at 105 (vis-à-vis the US Dollar), which will directly weigh on the country’s economic outlook as fundamentals will start to deteriorate, leaving less and less room for some BoJ manoeuvre.

Chart: Oil prices (WTI, 6M Lead) vs. Japan Trade Balance (Source: Eikon Reuters)

Japan Trade

Great Chart: US Term Premium vs. Business Cycles

Academics and economists have often decomposed the long-term bond yield of a specific country (i.e. US 10Y Treasury) into the sum of the expected path of real interest rate (r*) and the additional term premium, which compensate investors for holding interest rate risk. Two major risks that a bond investor typically face in the long-run are the change in supply of and demand for bonds and the uncertainty around inflation expectations. If the uncertainty increases, the market will demand a higher premium as a response. As the premium is not directly observable, it must be estimated using econometric models. For instance, a popular one that practitioners use is the one developed by Adrian, Crump and Moench (2013), who estimated fitted yields and the expected average short-term interest rates for different maturities (1 to 10 years, see data here).

As you can see, the term premium has been falling since 2009 and is currently negative at -51bps, which has not happened very often. Instead of having a positive term premium for long-term US debt holders carrying interest rate risk, there is actually a discount. The term premium for the 10Y reached an all-time low of -84bps in July 2016, at the same time that the yield on the Treasury reached a record low below 1.40%. However, there are also interest findings when we plot the ACM 10Y term premium with macroeconomic variables. If we overlay it with the US unemployment rate, we can notice a significant co-movement between the two times series. The jobless rate went down from 10% in Q3 2009 to 4.1% in March 2018, tracked by the term premium that fall from roughly 2.5% to -50bps in that same period. In other, it seems that the term premium follows the business cycles, trend lower in periods of positive growth and falling unemployment and rises in periods of contractions. Therefore, for those who are expecting a rise in the US LT yields in the medium term, driven by a reversion in the term premium, what does it mean for the unemployment rate going forward?

Chart. US Unemployment Rate vs. 10Y Term Premium

US Term Pr.png

Source: Reuters Eikon and Adrian et al. (2013)